<?xml version="1.0" encoding="UTF-8"?><rss version="2.0" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Robotics Daily · Learning and Making</title><description>Every Robotics Daily capsule in the segment Learning and Making, the latest first.</description><link>https://roboticsdaily.robotshop.com/</link><language>en</language><atom:link href="https://roboticsdaily.robotshop.com/feeds/learning-and-making.xml" rel="self" type="application/rss+xml"/><item><title>MABEL open-source bimanual robot is built from about $8.1k of parts</title><link>https://roboticsdaily.robotshop.com/capsules/nyu-mabel-open-bimanual-8056-2026-11/</link><guid isPermaLink="true">https://roboticsdaily.robotshop.com/capsules/nyu-mabel-open-bimanual-8056-2026-11/</guid><description>A preprint submitted on 2 November 2026 describes a 56-degree-of-freedom wheeled robot with two arms and two dexterous hands.</description><pubDate>Thu, 05 Nov 2026 00:00:00 GMT</pubDate><content:encoded>&lt;figure&gt;&lt;img src=&quot;https://roboticsdaily.robotshop.com/_astro/nyu-mabel-open-bimanual-8056-2026-11.BJ26kzxT_ywoXa.jpeg&quot; alt=&quot;A white MABEL bimanual robot with an orange-accented camera head, two jointed arms with dexterous tactile hands and a wheeled base stands in a lab hall, with orange test panels behind it.&quot; /&gt;&lt;figcaption&gt;Image: The MABEL Authors / project website · press-kit&lt;/figcaption&gt;&lt;/figure&gt;&lt;p&gt;A team including Ludovic Righetti submitted a preprint to arXiv on 2 November 2026 describing MABEL, an open mobile bimanual robot. Humanoids Daily attributes the work to the NYU Center for Robotics and Embodied Intelligence, which the arXiv abstract page does not state.&lt;/p&gt;&lt;p&gt;MABEL has 56 degrees of freedom, from two 7-degree arms, two 17-degree hands, a pitching torso, a lift, a 3-degree neck and a holonomic base. Its fingertips reach from floor level to 2.23 m. The authors put the parts cost at about US$8.1k.&lt;/p&gt;&lt;p&gt;The robot is teleoperated for the whole body with Apple Vision Pro head and hand tracking, without handheld controllers, with a mean glass-to-glass stereo video latency of 137 ms. The authors also demonstrate an end-to-end imitation-learning pipeline, from teleoperated data collection to policy deployment. The abstract reports no quantitative policy success rates, and the paper is a preprint that has not been peer reviewed.&lt;/p&gt;</content:encoded><category>Learning and Making</category></item><item><title>D-Robotics and Unitree launch an embodied-AI education package</title><link>https://roboticsdaily.robotshop.com/capsules/d-robotics-unitree-embodied-ai-education-2026/</link><guid isPermaLink="true">https://roboticsdaily.robotshop.com/capsules/d-robotics-unitree-embodied-ai-education-2026/</guid><description>D-Robotics and Unitree showed a joint embodied-intelligence curriculum for schools at the Changsha Education Equipment Exhibition, reported on 3 November 2026.</description><pubDate>Tue, 03 Nov 2026 00:00:00 GMT</pubDate><content:encoded>&lt;figure&gt;&lt;img src=&quot;https://roboticsdaily.robotshop.com/_astro/d-robotics-unitree-embodied-ai-education-2026.BHo2BHxR_Z1OgBNX.jpeg&quot; alt=&quot;A Unitree quadruped robot dog with a black robotic arm and sensor module mounted on its back reaches toward a stand of drink cans in a white demo arena at a D-Robotics exhibition booth.&quot; /&gt;&lt;figcaption&gt;Image: D-Robotics, via Gasgoo · press-kit&lt;/figcaption&gt;&lt;/figure&gt;&lt;p&gt;D-Robotics and Unitree Robotics made their first joint appearance at the Changsha Education Equipment Exhibition in China and launched an embodied-intelligence education package, Gasgoo reported on 3 November 2026. The article says the exhibition recently opened but gives no exact dates.&lt;/p&gt;&lt;p&gt;The package is built on D-Robotics&amp;#39; Sunrise S100 computing box, with Unitree&amp;#39;s Go2 quadruped and R1 humanoid as the larger platforms. Students begin on desktop robots, OriginBot and OriginMan, which run on the RDK X3 and X5 computing bases, then move to the Go2 and the R1.&lt;/p&gt;&lt;p&gt;The curriculum has 15 themed courses totalling 180 lessons, for primary through high school, including six courses linked to competitions. According to Gasgoo, the two companies plan to build further education solutions on the Sunrise S100 and then extend the partnership to D-Robotics&amp;#39; wider Sunrise S and X series for education, industrial and other uses.&lt;/p&gt;&lt;p&gt;The report is a trade-press account of the launch. It carries no price, availability date or school adoption figures, and no direct release from either company was found.&lt;/p&gt;</content:encoded><category>Learning and Making</category></item><item><title>OpenBDX, an open-source walking duck robot, has 25 builds under way</title><link>https://roboticsdaily.robotshop.com/capsules/openbdx-duck-robot-make-2026/</link><guid isPermaLink="true">https://roboticsdaily.robotshop.com/capsules/openbdx-duck-robot-make-2026/</guid><description>Make: profiles OpenBDX, an open build of a duck-style bipedal robot whose site has more than 1,300 verified users and at least 25 robots in progress.</description><pubDate>Sat, 31 Oct 2026 00:00:00 GMT</pubDate><content:encoded>&lt;figure&gt;&lt;img src=&quot;https://roboticsdaily.robotshop.com/_astro/openbdx-duck-robot-make-2026.Bbgk1bpW_Z1dkKqn.jpeg&quot; alt=&quot;A small bipedal duck-style robot with a grey head bar, two glowing round eyes, antennas and red-orange feet stands on a suburban sidewalk beside a lawn.&quot; /&gt;&lt;figcaption&gt;Image: Jaime Machuca / OpenBDX, via Make: · press-kit&lt;/figcaption&gt;&lt;/figure&gt;&lt;p&gt;OpenBDX, an open-source bipedal robot styled after a small walking duck, is the subject of a Make: article published on 30 October 2026 and written by the project&amp;#39;s founder, Denver engineer Jaime Machuca. Since openbdx.org opened on 2 September, it has drawn more than 1,300 registered, verified users, and at least 25 ducks are being built around the world. Builders sign in with GitHub to get the files and the build guide.&lt;/p&gt;&lt;p&gt;Each leg has five degrees of freedom, and the head and neck add four more plus two antenna servos; the walking policy drives 14 joints. Two Nvidia Jetson Orin NX modules, one in the chest and one in the head, share the work. Walking is learned with reinforcement learning in simulation, mainly MJLab with PPO, using animation references, pushes, rough terrain and domain randomization, and the policies are exported to ONNX to run on the robot. Mimir Reynisson led much of the hardware, firmware and training, and Florian Bieregger redesigned the CNC leg parts.&lt;/p&gt;&lt;p&gt;The article gives no build cost. The project shows how sim-to-real legged locomotion, once a research-lab capability, is now within reach of a hobbyist community.&lt;/p&gt;</content:encoded><category>Learning and Making</category></item><item><title>Opentrons raises $40 million to standardize AI-driven science on its lab robots</title><link>https://roboticsdaily.robotshop.com/capsules/opentrons-40m-series-e-ai-lab-robots-2026/</link><guid isPermaLink="true">https://roboticsdaily.robotshop.com/capsules/opentrons-40m-series-e-ai-lab-robots-2026/</guid><description>Opentrons announced a $40 million Series E on 29 October 2026, led by healthcare investor Avego, for its open lab automation robots.</description><pubDate>Fri, 30 Oct 2026 00:00:00 GMT</pubDate><content:encoded>&lt;figure&gt;&lt;img src=&quot;https://roboticsdaily.robotshop.com/_astro/opentrons-40m-series-e-ai-lab-robots-2026.DjPOhNy8_ZRRk5j.jpeg&quot; alt=&quot;A studio render of the Opentrons Flex lab robot, a black and silver enclosure with a glowing blue light bar, gantry pipetting head, gripper and touchscreen, shown over a labware deck against a clean white background.&quot; /&gt;&lt;figcaption&gt;Image: Opentrons (2023 Flex launch image, via TechCrunch) · press-kit&lt;/figcaption&gt;&lt;/figure&gt;&lt;p&gt;Opentrons announced a $40 million Series E round on 29 October 2026. The round was led by Avego, a healthcare investor, with participation from BroadOak Capital Partners. Existing investors named in the release include SoftBank Vision Fund and Khosla Ventures.&lt;/p&gt;&lt;p&gt;The company makes open, API-driven laboratory robots. It says more than 11,000 of them are deployed in over 80 countries, and that its customers include 90 percent of the top 20 global pharmaceutical companies, 14 of the top 15 global biopharma companies and every top-20 US research university. These are company claims.&lt;/p&gt;&lt;p&gt;Opentrons says the money will go to robotics, software and data infrastructure, AI-driven experiment execution, new automation capabilities, new markets and geographies, faster protocol development and deeper integration with the wider lab automation ecosystem.&lt;/p&gt;&lt;p&gt;The release does not give the total raised to date or a valuation. It describes the aim as standardizing biological experimentation on its robots, so that AI systems can run experiments in the physical lab.&lt;/p&gt;</content:encoded><category>Learning and Making</category></item><item><title>ETH Zurich designs lab-grown muscle actuators with up to 11 times the range of motion</title><link>https://roboticsdaily.robotshop.com/capsules/eth-zurich-biohybrid-muscle-robots-natcomm-2026-10/</link><guid isPermaLink="true">https://roboticsdaily.robotshop.com/capsules/eth-zurich-biohybrid-muscle-robots-natcomm-2026-10/</guid><description>ETH Zurich researchers reported in Nature Communications in October 2026 biohybrid actuators of lab-grown muscle with up to eleven times the range of motion of earlier systems.</description><pubDate>Tue, 27 Oct 2026 00:00:00 GMT</pubDate><content:encoded>&lt;figure&gt;&lt;img src=&quot;https://roboticsdaily.robotshop.com/_astro/eth-zurich-biohybrid-muscle-robots-natcomm-2026-10.Df-aHvKf_Zp3oIo.jpeg&quot; alt=&quot;Black-and-white close-up of a small biohybrid muscle robot arching in a water tank, labelled ON, with a 5 mm scale bar.&quot; /&gt;&lt;figcaption&gt;Image: ETH Zürich / YouTube · video-thumbnail&lt;/figcaption&gt;&lt;/figure&gt;&lt;p&gt;Researchers at ETH Zurich&amp;#39;s Soft Robotics Lab, led by Robert Katzschmann, reported in Nature Communications biohybrid actuators made of lab-grown mouse muscle, Robotics &amp;amp; Automation News reported on 23 October 2026. The team designed the muscle tissue and its hydrogel scaffold together rather than separately.&lt;/p&gt;&lt;p&gt;The resulting actuators have up to eleven times the range of motion of earlier systems with a comparable volume of muscle. About 1.5 centimetres long, they can jump, swim, crawl and grasp; a single contraction lifts one of them about 5 millimetres, and a walking version moves at 6.2 millimetres per second.&lt;/p&gt;&lt;p&gt;The figures come from the trade report, as the paper page could not be opened here. The researchers name an artificial blood supply as one of the main hurdles before such actuators can grow larger or last longer.&lt;/p&gt;</content:encoded><category>Learning and Making</category></item><item><title>MIT steers a paper-thin robot powered by muscle cells through a water maze</title><link>https://roboticsdaily.robotshop.com/capsules/mit-raman-muscle-film-biohybrid-swimmer-2026/</link><guid isPermaLink="true">https://roboticsdaily.robotshop.com/capsules/mit-raman-muscle-film-biohybrid-swimmer-2026/</guid><description>MIT researchers reported on 29 September 2026 a half-millimetre robot film carrying light-responsive muscle cells, steered by flashes of light through a maze.</description><pubDate>Tue, 27 Oct 2026 00:00:00 GMT</pubDate><content:encoded>&lt;figure&gt;&lt;img src=&quot;https://roboticsdaily.robotshop.com/_astro/mit-raman-muscle-film-biohybrid-swimmer-2026.CRFaKWNB_2ve3uc.jpeg&quot; alt=&quot;A gloved hand points a thin blue light probe at a petri dish on a grid mat; inside the dish, clear barriers form a small maze and blue light glows where the light hits the water.&quot; /&gt;&lt;figcaption&gt;Image: Melanie Gonick, MIT · press-kit&lt;/figcaption&gt;&lt;/figure&gt;&lt;p&gt;Researchers in Ritu Raman&amp;#39;s lab at MIT described on 29 September 2026, in the journal Advanced Functional Materials, a small swimming robot powered by living muscle. Its body is a half-millimetre-thin film of GelMA hydrogel carrying a single layer of skeletal muscle cells that contract in response to light, arranged on two fins.&lt;/p&gt;&lt;p&gt;Flashing light on one fin or the other makes the robot turn, which let the team steer it through a simple maze in water. Its top speed is about four body lengths per minute, well beyond the lab&amp;#39;s earlier iris-shaped design, which moved about 100 microns. The work was supported in part by the U.S. Office of Naval Research.&lt;/p&gt;&lt;p&gt;The robot is a laboratory demonstration with no practical application yet. Biohybrid robots are a research field in which universities test how living tissue can serve as an actuator, and the methods used here can be followed in a well-equipped lab.&lt;/p&gt;</content:encoded><category>Learning and Making</category></item></channel></rss>